Commentaries

Algos gone wild

The many proponents of high-frequency trading keep saying there’s no reason to be concerned about a rogue algorithim sparking a 1987 market-style crash. HFT supporters keep saying show us a case where a rogue algo even caused a minor hiccup in the market.

Well, Bernard Donefer, a professor at CUNY’s Baruch College in New York City and a critic of highly-automated trading programs, says the world already has gotten a glimpse at the kind of mayhem a rogue or simply a misfiring algo can cause.

Donefer, in a soon to be published research paper, blames high-frequency traders and an algo gone wild for a bizarre $9 drop in United Airlines’ stock on Sep. 8, 2008. The sudden plunge in UAL shares wiped out $1 billion in market value in just 12 minutes, after a six-year-old headline about the airline filing for bankruptcy erroneously hit some news wires.

The airline’s stock quickly recovered after it was determined that the bankruptcy story was an old, old story. But Donefer argues the percipitous drop in UAL shares “was mostly the result of the interplay between the algorithms that search and compile information from the Web and the ones that Wall Street firms and hedge funds use to make trades automatically.”

This is an isolated case, but Donefer says it’s only a matter of time before an event like the UAl one–or a series of events in which algos go wild–sparks a widespread market crash.

Will we see an event caused by algos gone wile in our markets? I believe it is inevitable. I am further convinced that with no planning…or regulatory framework it will be hard to stop. With unfettered or naked access, it might impact the viability of a broker.

This is the doomsday scenario I wrote about in my column Wall Street meets The Matrix. It’s also the kind of computer-driven catastrophe that the folks at Zerohedge.com and Joe Saluzzi of Themis Trading have been warning about.

Wall Street meets The Matrix

Michael Durbin is no Wall Street rebel. But Durbin, who has been on the front lines of high-frequency trading (HFT) since its early days, isn’t afraid to buck the industry line that lightning-fast trading of stock, options and commodities poses little or no risk to the stability of the markets.

Durbin says it’s reasonable to wonder whether Wall Street’s unfettered embrace of algorithmic automated trading could be setting the stage for a future meltdown.

“You have multiple HFT trading firms and sometimes their agendas are complementary and sometimes they’re not,” explains Durbin, director of HFT research with Blue Capital Group, a small Chicago-based options trading firm.

“There could be a time where these HFT programs unintentionally collaborate and you have a two- or three-minute period where the markets are going crazy. Then other traders respond to it and it simply gets out of control.”

What Durbin’s talking about is the dreaded contagion effect, in which a bad trade or a rogue algorithm misfires — sparking copycat sell orders at other high frequency desks.

It’s the kind of machine-driven crash that sounds like the plot line for “Wall Street” meets “The Matrix”.

High frequency trading programs are designed to scour the markets to decipher trends in trading patterns and place buy and sell orders a millisecond ahead of the pack. It all happens at warp speed, and except for developing the algorithmic programs, the human element is all but gone from the equation.